Machine Learning Analysis of Enhanced Biodegradable Phoenix dactylifera L./HDPE Composite Thermograms

Area of Science:

• Materials Science
• Polymer Chemistry
• Computational Science

Background:

• Growing focus on waste recycling and decomposition for economic value and resource management.
• Biodegradable polymer composites offer potential for sustainable material solutions.
• Phoenix dactylifera L. (PD) and high-density polyethylene (HDPE) are abundant waste materials suitable for composite development.

Purpose of the Study:

• To explore the thermal decomposition of enhanced biodegradable polymer matrices (PD/HDPE).
• To apply machine learning analysis to experimental thermogravimetric analysis (TGA) data.
• To develop and validate predictive models for PD/HDPE thermal behavior.

Main Methods:

• Thermogravimetric analysis (TGA) of PD/HDPE composites at varying heating rates (10-40 °C·min⁻¹) and temperatures (25-600 °C).
• Development of a deep neural network machine learning model.
• Training and validation of learning algorithms with experimental TGA data.

Main Results:

• TGA showed weight loss dependent on degradation temperature, with thermograms shifting at higher heating rates.
• PD/HDPE composites exhibited significant residual ash, with HDPE being more thermally stable than PD.
• Machine learning model achieved near-zero residual error (0.025) and perfect correlation (R²~1) with experimental data.

Conclusions:

• Machine learning accurately predicts the thermal decomposition of PD/HDPE composites.
• Optimized algorithms and models reduce the need for extensive experimental testing.
• This approach enables efficient prediction and optimization of thermal characteristics for waste-derived polymers.